17 Our methods included semistructured in-depth selleck compound interviews (see interview guides, online supplementary appendix 1) with policymakers (N=12), administrators (N=43), TCAM practitioners (N=59), allopathic practitioners (N=37), traditional healers (N=7), as well as health workers and community representatives (N=38) in three diverse Indian states (figure 1). We undertook the study in Kerala, where a number of systems have strong historical and systemic roots (N=74); Meghalaya, where local health traditions hold sway (N=61); and Delhi, where national, state and municipal jurisdictions
interface with multiple systems of medicine (N=61). Participants were selected based on maximum variation criteria for each category. We sought to represent different schemes, levels of implementation (directorates, zonal officers), systems of medicine, types of establishments (hospital, dispensary) and years of experience. Figure 1 Location of states in India where fieldwork was conducted (New Delhi, Meghalaya and Kerala). In each state, one senior researcher, a research associate and a field researcher developed selection matrices to achieve maximum variation across each category of respondents.
In each state, districts (two in Kerala, and three in Meghalaya) or municipal zones (three in Delhi) were chosen
to represent variation in accessibility and distribution across TCA providers (eg, small or large proportions of local health practitioners, Homoeopaths, Ayurvedic and/or Unani practitioners). Publicly available records from state and municipal authorities were consulted in order to determine the location and type of facility (co-located, stand-alone) as well as suggestions and recommendations from key informants. We also ensured that facilities closest to and furthest from district headquarters were chosen for interviews, to maximise variability. We would typically contact providers via cell phone, share information about the study verbally or via email, and set up a time to interview them in-person. In some cases, we would arrive during outpatient clinic hours at the chosen facility, share our participant information sheet and seek an appointment Carfilzomib time with eligible participants. In most cases, we found that participants were keen to participate once they were aware of the nature of the study and, in some cases, the assurance of confidentiality. We had no refusals, although some allopathic practitioners had to be persuaded to participate by emphasising that this study was not ‘pro-TCAM integration’ per se, but merely seeking to understand state policy implementation.